automatic vacuum on pg_statistic pg_toast area blocks all queries in hot standby

Hi,

We have a master/hot standy databases set up with streaming replication. The
pair are running:

PostgreSQL 9.1.11 on x86_64-unknown-linux-gnu, compiled by gcc (GCC) 4.4.7
20120313 (Red Hat 4.4.7-3), 64-bit
Linux 2.6.32-431.11.2.el6.x86_64 #1 SMP Tue Mar 25 19:59:55 UTC 2014 x86_64
x86_64 x86_64 GNU/Linux

We just find the following situation:

on the master db server:
(1) at 16:58:22 automatic vacuum of table ""pg_toast.pg_toast_2619"": index
scans: 1

almost at the same time, queries at hot standby are blocked, all waiting for
a share lock on pg_toast_2619, for example:

process 26822 still waiting for AccessShareLock on relation 2840 of database
16414 after 1000.089 ms
process 26820 still waiting for AccessShareLock on relation 2840 of database
16414 after 1000.085 ms
process 26819 still waiting for AccessShareLock on relation 2840 of database
16414 after 1000.079 ms
and etc...


relation 2840 is actually pg_toast_2619 of pg_statistic:

DB=# select relname from pg_class where reltoastrelid=2840;
   relname
--------------
 pg_statistic
(1 row)


Any idea why?

Thanks a lot,

Pius




--
View this message in context:
http://postgresql.1045698.n5.nabble.com/automatic-vacuum-on-pg-statistic-pg-toast-area-blocks-all-queries-in-hot-standby-tp5807143.html
Sent from the PostgreSQL - bugs mailing list archive at Nabble.com.